The nature of categories and their acquisition is one of the central open questions in Cognitive Science. We suggest that categories are represented via structured descriptions and formed by a process of progressive abstraction, through successive comparison with incoming exemplars. This paper describes how SEQL (Skorstad, Gentner, & Medin, 1988), a computer model for category learning, which is based on SME (Falkenhainer et al 1986, 1989; Forbus et al 1994) can be used to simulate a recent categorization experiment (Ramscar & Pain, 1996), using a new algorithm, Generalization and Exemplar Learning (GEL). We demonstrate that SEQL produces behavior consistent with human subjects. Introduction Similarity is often viewed as central to categorization. For instance, prototype theories of categorization posit that categorization decisions are made on the basis of the similarity of an entity to the prototypical member of that category (Rosch 1975). However, similarity-based accou...

Abstract not found

. Whilst many user models can function perfectly adequately with a behavioural impression of the user, the provision of assistance in some task domains, notably design, requires a richer understanding, incorporating information about the user's knowledge and beliefs. This raises a number of important and difficult questions: How can we know what the user knows, and how can we know that we know? We present evidence that the psychological view of human conceptual knowledge that underpins typical approaches to these questions is flawed. We argue that user knowledge can be modelled, up to a point, but that to ask whether or not we can know what the user knows is to misunderstand the question. 1 What Do We Want to Know? Many user models can function perfectly adequately with only a behavioural impression of the user: A user's actions can be sufficient input for a system to adapt in order to accommodate the particular needs of a particular user or user type. Even if we include a user's li...